Underwater AUV swarms can explore large regions and search for multiple targets in parallel but the limited range of acoustic links and the absence of infrastructure make decentralized self-allocation challenging. This paper presents the Ad hoc network Assisted Firefly system (AA-FY), a firefly-inspired method integrating sensing, multi hop underwater ad hoc communication, and movement. It bridges Multi-Robot Task Allocation (MRTA) and Multimodel Optimization (MMO) for a robotic search system, where MMO is a theoretical guide solving an MRTA problem. System-wise, AA-FY applies a bio-inspired movement strategy to accomplish the search and self-allocation for aggregations at multiple targets. The ad hoc strategy allows information to be aggregated and disseminated only when superior information arrives. It also guides the movement using an attenuation clue based on hop counts. Such strategies encourage localized convergence and stable niches around multiple targets. Further, AA-FY uses a greedy selection strategy to drive each AUV toward the most influential attenuated candidate without a central controller or global map. The evaluations show that AA-FY reliably allocates clusters of AUVs around all targets. It even succeeds when targets are far apart or spatially biased. AA-FY also achieves proportional allocation according to target signal strength.
Yang et al. (Fri,) studied this question.